For v2rayN users dealing with core startup failures, node timeouts, or websites that won't load. Using the Xray core on Windows as an example, capture the logs for a single failed attempt, then use the error address and operation to distinguish local listening issues, DNS lookup failures, remote connection problems, and TLS handshake errors. Change one thing at a time and test again.
Open the log and capture a failed attempt
Start with the log panel in the v2rayN main window. It may show both client activity and core output: a subscription update failure describes fetching a subscription URL, while a core dial failure describes a particular network connection. Even if the messages appear next to each other, they may not belong to the same request.
Keep test conditions consistent
Keep the current node, routing setup, and connection method unchanged, and note whether the system proxy or TUN is enabled. If startup fails, reproduce only the startup failure. If a website won't load, refresh a single page that normally works. Don't run a speed test or update subscriptions at the same time.
Check the log level
Go to “Settings” → “Parameter Settings” and find the log level under the core's basic settings. For Xray,
warningis a good starting point for checking reported warnings and errors; switch toinfowhen you need to see how connections are handled. The setting's location varies by client version, so first confirm which core the current connection uses.Apply the setting and reproduce the issue
Save the setting and make sure the core has reloaded its configuration. Note the time, then reproduce the issue once. Capture the lines from that time window that show the request and its failure reason. If the lines include a request ID, match entries with the same ID to avoid confusing a background request with the issue you're investigating.
Keep the key details
Note whether the failure occurred during
listen,lookup,dial, or certificate verification. Before sharing logs, redact subscription URLs, user identifiers, real domain names, and public IP addresses. Keep the port, error text, and whether each address is local or remote.
Common errors: timeouts, connection refused, and handshake failures
The errors below are searchable snippets, not complete messages that every core will produce. If a line chains together several errors, find the underlying cause that includes a specific address or failed operation first, then check which inbound, outbound, or request produced the outer error.
Error: i/o timeout
Cause: A network operation did not complete before its deadline. It could be connecting to a remote host, reading data, or waiting for a DNS lookup.
Fix: If the same line says dial tcp, check the node's remote host and port against the address in the log, then test TCP connectivity. If it shows lookup, investigate the DNS resolution path first.
Error: no such host
Cause: The system making the query couldn't resolve the hostname. The name that failed may not be the website open in your browser.
Fix: Check the full hostname after lookup and compare it with the node configuration. Then compare system DNS resolution with the DNS path used by the core; a successful system lookup alone doesn't rule out a core DNS issue.
Error: connection refused
Cause: The connection was rejected, often because nothing is listening on the target port, though an intermediary device may also reject it.
Fix: If the target is 127.0.0.1, compare the port used by the application with the port the core listens on. For a remote target, check the port in the subscription and ask the service provider whether it is listening.
Error: x509: certificate signed by unknown authority
Cause: The TLS connection reached certificate verification, but its trust chain could not validate the peer's certificate.
Fix: First determine whether the TLS peer in the log is the node or the destination website. Then check the certificate chain and local trust settings. Don't bypass verification to suppress the error.
Error: remote error: tls: handshake failure
Cause: The remote server returned a failure alert during the TLS handshake. This generic error alone doesn't reveal which setting is mismatched.
Fix: For a node using TLS, compare the transport type, TLS setting, and server name (SNI) against the server's configuration. If needed, ask the service provider to check the server logs for the same time.
Local port conflicts: identify the listener before changing ports
If core startup reports address already in use or the Windows error Only one usage of each socket address, note the address and port that failed to bind. The two ports below illustrate different locations in a log; when troubleshooting, follow your own settings and error message.
Determine which side to change
If the error is listen tcp 127.0.0.1:10808, check what's listening locally; don't change the remote port in the subscription. If a dial tcp connection to remote port 443 times out, changing the local port won't restore the remote connection.
In Windows PowerShell, check which process is listening on a TCP port. Replace the number on the first line with the port that failed to bind in the log. If the log explicitly says UDP, use Get-NetUDPEndpoint -LocalPort $port to check the corresponding endpoint.
$port = 10808
Get-NetTCPConnection -LocalPort $port -State Listen |
Select-Object LocalAddress, LocalPort, OwningProcess
- Compare
OwningProcesswith the PID on the Task Manager “Details” tab. If the listening address is0.0.0.0, also check whether it conflicts with the local address the core is trying to bind to. - If the port is occupied by a duplicate client instance, exit the extra instance normally from the system tray. If it's used by a program that needs to stay open, change the corresponding local listening port in v2rayN under “Settings” → “Parameter Settings,” and update any applications manually configured to use that port.
- Restart the core and check the listening status again. Confirm that the process has bound to the expected port, then try the website again. If the port is listening but the page still won't load, check the inbound and outbound logs for the new request instead of repeatedly changing ports.
Core is running: check traffic capture, routing, and remote connectivity
A successful core startup only means the process can run. It doesn't mean the browser's request reached the core, or that the node or destination website is reachable. Temporarily set the log level to info, reproduce the issue with the same URL and connection method, and check these three points.
- Check inbound traffic. If there's no connection entry for the attempt, first check whether the application is using the system proxy or manually pointing to the correct local port. With TUN, check that it started successfully and that the traffic is going through it. Verify the current connection method before switching modes; don't change both at once.
- Check outbound routing. Once an inbound entry appears, compare the outbound tag in the log with the active routing rules. If a domain expected to go through a node matched a direct rule, check the rule match and order first instead of changing the node's certificate settings.
- Check the remote connection. If the request used the expected outbound and failed while connecting to a TCP node, use PowerShell's
Test-NetConnectionto test the node's hostname and actual port. This only checks whether a TCP connection can be established from the current environment. It doesn't verify TLS, protocol authentication, or website availability, and it can't determine whether UDP transport works.
Retest with the same node and URL after each change, and compare logs from the same stage. If a dial timeout turns into a certificate error, troubleshooting has moved to the next layer; it doesn't mean the connection is fixed. Once the website loads again, restore the original log level.
v2rayN log troubleshooting FAQ
The log is scrolling too fast. How do I find the entries for my request?
Pause speed tests and subscription updates, reproduce the issue once, and note when you started. Narrow down the log by time, then use the destination address or request ID to link related entries. Switch briefly to debug only if info still lacks key context, and restore the original level after collecting the logs.
Why did the subscription update time out when websites still load?
A subscription update connects to the subscription server, possibly using a different connection method from your browser. Update only the group that failed and note the error. Then check whether the subscription URL is still valid, whether the update uses a proxy, and whether the node works at the same time. Don't delete existing groups yet.
The log says the certificate has expired. Do I need to replace the node?
First, check and sync the system time in Windows under “Settings” → “Time & language” → “Date & time,” then test again. If the time is correct and the certificate error persists, identify the TLS peer in the log and ask the party responsible for it to check the certificate's validity period and deployment.
Why won't a website load even though the speed test shows a result?
The speed test and the website request are separate connections. Stop the test and visit the site on its own. In the new logs, check which outbound the request used and whether it failed during DNS resolution, dialing, or the handshake. If only this domain fails, first check its routing rule match and DNS result.